Fuzzy clustering and fuzzy sets for reliability analysis of recent distribution systems

Abstract

This paper presents the application of fuzzy clustering to reduce the large system states into few representative clusters, which are sufficient for reliability analysis of recent distribution networks containing local photovoltaic generation. For each cluster the failure effect analysis and minimal cut-set techniques, based on outage simulation of system components, are applied to compute the different reliability indices. The uncertainties of failure and repair rates for the components of the obtained cut sets are taken into consideration by applying fuzzy set theory. The computed indices are then multiplied by the corresponding cluster weight to obtain the overall system indices. The developed algorithm is applied to a 66-kV distribution network and the resulting improvement of system reliability is estimated and discussed in the paper.